Right Cargo, Wrong Address

In vivo gene therapy — correcting or replacing genes directly inside the patient's body — requires delivery vectors that simultaneously cross biological barriers (blood vessel walls, cell membranes, endosomal escape, nuclear entry), target the correct cell type with high specificity, carry sufficient genetic payload, and evade the immune system. No delivery vector achieves all four requirements. Viral vectors (AAVs) achieve good cell entry but trigger neutralizing antibodies that prevent repeat dosing, carry limited payload (<5 kb for AAV), and have tropism that is difficult to redirect to arbitrary cell types. Non-viral vectors (lipid nanoparticles) naturally accumulate in the liver and achieve less than 1% delivery efficiency to non-liver tissues.

Over 10,000 human diseases have identified genetic causes, but only ~30 gene therapies have been approved — almost all for rare diseases affecting small patient populations. The current dominant approach is ex vivo therapy (removing cells, editing them outside the body, and reinfusing), which costs $1–3.5 million per treatment. In vivo delivery would eliminate the need for cell harvesting and reinfusion, potentially reducing costs by 10–100× and making gene therapy accessible for common diseases (e.g., sickle cell disease affects 20+ million people globally). Without solving the delivery problem, gene therapy will remain limited to rare diseases treatable by liver-targeted or ex vivo approaches.

Adeno-associated viruses (AAVs) are the most clinically advanced vectors but face three limitations: pre-existing neutralizing antibodies in 30–60% of the population (depending on serotype) prevent treatment; redosing triggers immune responses; and the 4.7 kb packaging capacity excludes large genes like dystrophin (14 kb, needed for Duchenne muscular dystrophy). Lipid nanoparticles (LNPs) — the technology behind mRNA COVID vaccines — work well for liver targeting but >90% of injected dose accumulates in the liver regardless of surface modifications. Redirecting LNPs to lung, brain, or muscle tissue has been attempted via ligand conjugation and charge modification, but efficacy in non-liver tissues remains below therapeutic thresholds. CRISPR delivery in vivo faces the additional challenge of delivering both the Cas protein and guide RNA to the same cell simultaneously.

Engineered AAV capsid variants (discovered through directed evolution or machine learning-guided design) that evade pre-existing antibodies and exhibit programmable tissue tropism would open the field. For non-viral approaches, understanding the biophysical mechanisms of LNP endosomal escape — currently estimated at <2% of internalized particles — could enable rational design of escape-enhancing lipid compositions. A "modular delivery platform" where targeting, immune evasion, and payload release components can be independently optimized and combined would accelerate progress across disease targets.

A student team could use computational approaches (molecular dynamics simulation of lipid-membrane interactions, or machine learning models trained on existing capsid-tropism datasets) to predict which structural modifications to AAV capsids or LNP lipid compositions would improve tissue-specific delivery. Alternatively, teams could design and test novel LNP formulations optimized for non-liver targeting using in vitro cell-type-specific uptake assays with fluorescent reporters. Relevant disciplines: biomedical engineering, chemical engineering, molecular biology, computational biology.